-
Expert-Augmented Machine Learning
Authors:
E. D. Gennatas,
J. H. Friedman,
L. H. Ungar,
R. Pirracchio,
E. Eaton,
L. Reichman,
Y. Interian,
C. B. Simone,
A. Auerbach,
E. Delgado,
M. J. Van der Laan,
T. D. Solberg,
G. Valdes
Abstract:
Machine Learning is proving invaluable across disciplines. However, its success is often limited by the quality and quantity of available data, while its adoption by the level of trust that models afford users. Human vs. machine performance is commonly compared empirically to decide whether a certain task should be performed by a computer or an expert. In reality, the optimal learning strategy may…
▽ More
Machine Learning is proving invaluable across disciplines. However, its success is often limited by the quality and quantity of available data, while its adoption by the level of trust that models afford users. Human vs. machine performance is commonly compared empirically to decide whether a certain task should be performed by a computer or an expert. In reality, the optimal learning strategy may involve combining the complementary strengths of man and machine. Here we present Expert-Augmented Machine Learning (EAML), an automated method that guides the extraction of expert knowledge and its integration into machine-learned models. We use a large dataset of intensive care patient data to predict mortality and show that we can extract expert knowledge using an online platform, help reveal hidden confounders, improve generalizability on a different population and learn using less data. EAML presents a novel framework for high performance and dependable machine learning in critical applications.
△ Less
Submitted 5 January, 2021; v1 submitted 22 March, 2019;
originally announced March 2019.
-
Tree-Structured Boosting: Connections Between Gradient Boosted Stumps and Full Decision Trees
Authors:
José Marcio Luna,
Eric Eaton,
Lyle H. Ungar,
Eric Diffenderfer,
Shane T. Jensen,
Efstathios D. Gennatas,
Mateo Wirth,
Charles B. Simone II,
Timothy D. Solberg,
Gilmer Valdes
Abstract:
Additive models, such as produced by gradient boosting, and full interaction models, such as classification and regression trees (CART), are widely used algorithms that have been investigated largely in isolation. We show that these models exist along a spectrum, revealing never-before-known connections between these two approaches. This paper introduces a novel technique called tree-structured bo…
▽ More
Additive models, such as produced by gradient boosting, and full interaction models, such as classification and regression trees (CART), are widely used algorithms that have been investigated largely in isolation. We show that these models exist along a spectrum, revealing never-before-known connections between these two approaches. This paper introduces a novel technique called tree-structured boosting for creating a single decision tree, and shows that this method can produce models equivalent to CART or gradient boosted stumps at the extremes by varying a single parameter. Although tree-structured boosting is designed primarily to provide both the model interpretability and predictive performance needed for high-stake applications like medicine, it also can produce decision trees represented by hybrid models between CART and boosted stumps that can outperform either of these approaches.
△ Less
Submitted 17 November, 2017;
originally announced November 2017.
-
Appropriation as neglected practice in communities: presenting a framework to enable EUD design for CoPs
Authors:
Federico Cabitza,
Carla Simone
Abstract:
Communities present considerable challenges for the design and application of supportive information technology (IT), especially when these develop in loosely-integrated, informal and scarcely organized contexts, like it is often the case of Communities of Practice (CoP). An approach that actively supports user communities in the process of IT appropriation can help alleviate the impossibility of…
▽ More
Communities present considerable challenges for the design and application of supportive information technology (IT), especially when these develop in loosely-integrated, informal and scarcely organized contexts, like it is often the case of Communities of Practice (CoP). An approach that actively supports user communities in the process of IT appropriation can help alleviate the impossibility of the members of these communities to rely on professional support, and enable even complex forms of tailoring and End-User Development (EUD). Although this approach has been already explored by an increasing number of researchers, however there is still a lack of a general framework that could play a role in the comparison of existing proposals and in the development of new EUD solutions for CoPs. The paper proposes a conceptual framework and a related architecture, called Logic of Bricolage, that aims to be a step further in this direction to enable better EUD-oriented support for digitized communities. The framework is described and the architecture instantiated in three concrete EUD environments that specifically regard collaborative activities in order to show the generality and applicability of the framework.
△ Less
Submitted 7 June, 2013;
originally announced June 2013.
-
Design Ltd.: Renovated Myths for the Development of Socially Embedded Technologies
Authors:
Federico Cabitza,
Carla Simone
Abstract:
This paper argues that traditional and mainstream mythologies, which have been continually told within the Information Technology domain among designers and advocators of conceptual modelling since the 1960s in different fields of computing sciences, could now be renovated or substituted in the mould of more recent discourses about performativity, complexity and end-user creativity that have been…
▽ More
This paper argues that traditional and mainstream mythologies, which have been continually told within the Information Technology domain among designers and advocators of conceptual modelling since the 1960s in different fields of computing sciences, could now be renovated or substituted in the mould of more recent discourses about performativity, complexity and end-user creativity that have been constructed across different fields in the meanwhile. In the paper, it is submitted that these discourses could motivate IT professionals in undertaking alternative approaches toward the co-construction of socio-technical systems, i.e., social settings where humans cooperate to reach common goals by means of mediating computational tools. The authors advocate further discussion about and consolidation of some concepts in design research, design practice and more generally Information Technology (IT) development, like those of: task-artifact entanglement, universatility (sic) of End-User Development (EUD) environments, bricolant/bricoleur end-user, logic of bricolage, maieuta-designers (sic), and laissez-faire method to socio-technical construction. Points backing these and similar concepts are made to promote further discussion on the need to rethink the main assumptions underlying IT design and development some fifty years later the coming of age of software and modern IT in the organizational domain.
△ Less
Submitted 14 March, 2013; v1 submitted 23 November, 2012;
originally announced November 2012.
